Accurate identification of biased measurements under serial correlation

Chemical process data are often correlated over time (i.e., auto or seriall y correlated) due to recycle loops, large material inventories, sampling la g, dead time, and process dynamics created by high-order systems and transp ortation lag. However, many approaches that attempt to identify gross error s in measured process variables have not addressed the issue of serial corr elation which can lead to large inaccuracies in identifying biased measured variables. Hence, this work extends the unbiased estimation technique (UBE T) of Rollins and Davis(1) to address serial correlation. The serially corr elated gross error detection study of Kao et al.(2) is used as a basis for setting up this study and comparison. In their work, the type of autocorrel ation was assumed known (ARMA(1,1)), and the measurement test (TWT) was use d for the identification of the measurement bias. While Kao et al.(2) used prewhitening of the data and variances of measured variables derived from k nowledge of the time correlation structure, this work presents two prewhite ning methods and a different identification strategy based on the UBET. Res ults of the simulation study show the UBET has higher perfect identificatio n rates and lower type I error rates over the MT.